Virtual Power Plant in South Australia · VPP-SA Milestone 2 Public Report 15-06-18 V1.1.docx 3 1 Introduction 1.1 Project overview AGL’s Virtual Power Plant (VPP) remains a key

Virtual Power Plant in South Australia Stage 2 Public Report

VPP-SA Milestone 2 Public Report 15-06-18 V1.1.docx 1

AGL Energy Limited

ABN: 74 115 061 375

Level 24, 200 George St

Sydney NSW 2000

Locked Bag 1837

St Leonards NSW 2065

t: 02 9921 2999

f: 02 9921 2552

agl.com.au

Table of Contents

Executive Summary ............................................................................................................ 2

1 Introduction ...................................................................................................................... 3

1.1 Project overview .................................................................................................................................3

1.2 Review of Stage 2 Objectives ...........................................................................................................5

2 Current Stage Review ............................................................................................. 6

2.1 Inclusion of new ESS technology ....................................................................................................6

2.2 Technology selection process..........................................................................................................6

2.3 Technology Announcement ..............................................................................................................7

2.6 Sales and installation metrics ..........................................................................................................9

2.7 VPP Performance ............................................................................................................................ 10

2.7.2 Fleet Dispatch Details and Results ............................................................................................... 11

2.7.3 Key Learnings ................................................................................................................................. 13

2.7.4 Day-to-Day Operation ..................................................................................................................... 14

2.7.5 Reduced Solar Export ..................................................................................................................... 16

2.7.6 VPP Performance Summary........................................................................................................... 17

2.8 Summary of Project Media ............................................................................................................. 17


Executive Summary

This milestone report describes AGL’s progress against the Stage 2 objectives of the 5MW Virtual Power

Plant in South Australia and is submitted in partial fulfilment of AGL’s knowledge sharing obligations to the

project’s funding partner, ARENA.

The project comprises the installation and orchestration of a 5MW VPP consisting of up to 1,000 residential

energy storage systems installed behind the meter, and capable of dispatching up to 12MWh of stored

energy. The project was established to demonstrate the role of virtual power plants in enabling higher

penetrations of distributed renewable generation in the grid and help further understanding about the role of

distributed ‘smart’ storage in supporting grid resilience and reliability in South Australia.

AGL has successfully completed Stage 2 of the three-stage deployment of energy storage systems. As at

31 May, 312 batteries had been installed in customer’s homes and the rollout of new technology energy

storage systems is progressing to a revised deployment plan considering hardware stock availability and

certified installation resources. Initial trials have demonstrated that the VPP can respond as expected to both

planned and unplanned dispatch events and has the potential to respond rapidly enough to participate in the

6 second contingency FCAS market.

Stage 2 Achievements

• Installations – more than 300 energy storage systems have been installed in customer premises.

• Sales – more than 700 batteries have been sold, and more than 500 requotes have been completed,

where customers chose to upgrade to the newest technology available as a part of the program. Lead

generation continues for the final tranche of customers with more than 800 leads generated. This has

been assisted by news coverage and energy storage media coverage particularly in the lead up to the

South Australian election in March 2018.

• Delivery and installation – AGL has seven trained installation crews and is ramping this up to ten crews.

AGL has the capacity to install up to 20 batteries per week and this will increase further as the project

continues.

• Stakeholder Reference Group – Three successful Stakeholder Reference Group (SRG) meetings have

been held.

• Knowledge Sharing – Presentations on project progress have been made to, among others, the

Australian Energy Storage Conference and NEM Future Forum.

• VPP Performance – The project has conducted demonstration dispatch events during 2017 and 2018

and has monitored the impact of the VPP on network load and site exports as well as the impact of local

grid conditions on individual energy storage systems and fleet performance.

Next Steps

Stage 3 of the project will focus on completing the remainder of the 1,000 installations and further

demonstration of VPP functionality for a number of network service and wholesale participation use cases.


1 Introduction

1.1 Project overview

AGL’s Virtual Power Plant (VPP) remains a key demonstration project as Australia’s energy markets continue

to evolve. The business model that underpins the VPP presents important opportunities for delivering more

reliable energy supply, empowering customers and contributing to the national policy debate regarding the

National Energy Guarantee (NEG).

As the cost of distributed energy resources (DER), including energy storage systems, continues to decline,

the increasing accessibility of these technologies enables traditional energy market participants and new

entrants to provide a range of new energy services offerings. In addition to allowing customers to more

actively manage their own energy needs, DER are also giving rise to new kinds of demand-side participation,

enabling customers to contribute system-wide benefits through dynamic programs. The once linear supply

chain is becoming increasingly decentralised and bi-directional. A more efficient deployment and use of DER

will enable co-optimisation across these multiple uses and value streams, including the management of

broader system needs.

AGL sees competition and innovation in technology and business models as the primary means for meeting

this co-optimisation challenge and aligning the interests of energy service providers with those of the

customers they serve. AGL’s VPP presents one such model, demonstrating at a commercial scale the

potential to orchestrate multiple DER to provide network services and wholesale market participation.

As the national policy debate regarding the NEG settles, AGL anticipates that retailers will look to develop a

range of innovative products and business models to meet their reliability obligation at least cost. Business

models that entail a behind-the-meter orchestration function, such as a VPP, could play an important role in

enabling dispatchable capacity when the system needs it.

AGL’s VPP seeks to demonstrate the value that grid-connected batteries can create for a range of

stakeholders when managed as part of a co-ordinated program. Stage 2 has seen the introduction of updated

energy storage technology and further demonstration of wholesale market dispatch events.

This report focuses on description of learnings from ESS installations and additional demonstration events,

and as well as an in-depth analysis of voltage observations.

Since AGL’s last milestone report, AGL’s VPP has faced several hurdles. High voltage levels in many parts

of the distribution network regularly affect some customers by making their ESS systems inoperable. During

voltage excursions, customers’ inverters disconnect from the grid, making them unavailable to the VPP. At

times when customers are exporting power into the grid, voltage levels are increasing further. Policymakers

and network businesses alike will need to consider these issues to facilitate the viability of the virtual power

plant model into the future.

AGL has now completed Stage 2 of its South Australian VPP, with the sale of over 700 and the installation

of over 300 Energy Storage Systems (ESS) as at 31 May 2018.


The following value streams have also been demonstrated using the VPP:

Value Stream Recipient Status

Increasing solar self-consumption Consumer Functionality demonstrated through ongoing performance

monitoring.

Utilising backup power in the event of an

outage Consumer

New technology now being rolled out has increased backup

functionality now being used by customers.

Network support through peak demand

management

Network service

provider Capability demonstrated through VPP activation events.

Network support through frequency control

services

Network service

provider Capability demonstrated through VPP activation events.

Physical hedge or arbitrage opportunity in

wholesale energy market Retailer Capability demonstrated through VPP activation events.


1.2 Review of Stage 2 Objectives

The primary objectives for Stage 2 were to complete the installation and commissioning of 300 Energy

Storage Systems (ESS) and to demonstrate the use of these as a VPP. These objectives have been

achieved.

Overall project performance against milestones has been as shown below:

Stage Installation

target

Completion

date

Completed

installations

Achievement

Date

Milestone 1 65 28 Feb 2017 150 30 Apr 2017

Milestone 2 300 31 May 2018 312 31 May 2018

Milestone 3 1,000 (min. 950) 31 Dec 2018

As at 31 May, a total of 312 ESS were installed and working. Test dispatch events were undertaken through

2017 and 2018 and the results of these are discussed in section 2.7.

There were also several secondary objectives in Stage 2. A review of the program’s performance against

these is as follows:

Objective Result

Consider integration of new energy storage systems

from suppliers other than Sunverge, if Recipient

determines technically and commercially feasible.

Tesla and SolarEdge/LG Chem are now included in the program.

Begin targeted deployment of energy storage The rollout of Tesla and SolarEdge/LG Chem systems in the field is

now well advanced.

Begin targeted deployment of energy storage systems in

specific network areas

Deployment of ESS to small geographic areas in the quantities

required to deliver network benefits has proved challenging. This is

discussed further in section Error! Reference source not found..

Refine targeted marketing channel Marketing channels now well-defined and working.

Bed down operational systems and focus on driving

down operating cost

Significant improvements have been made in the operational fulfilment of ESS resulting in an increase in deployment capacity from approximately 5 ESS per week to up to 20 per week.

A significantly lower cost of installation due to smaller form factor

batteries has also been achieved in Stage 2.

Demonstrate more sophisticated VPP capability VPP capability has been demonstrated.

Consideration and planning of Stage 3 technology

options to determine technical and commercial feasibility

Well advanced on the selection of new VPP orchestration software for

use during Stage 3.


2 Current Stage Review

AGL has now completed more than 300 ESS installations and has sold a further 400 ESS units that are being

deployed in customer’s homes at a rate of 15 - 20 units a week.

2.1 Inclusion of new ESS technology

In March 2018, AGL added two new technologies to the SA-VPP program. Following a thorough selection

and technical due diligence process, Tesla’s Powerwall 2, and LG Chem’s RESU10H batteries paired with a

SolarEdge inverter were made available to customers as part of a next generation technology offering.

These new technologies reflect advancements in the market since the program was launched. The new

products provide customers with up to 35% more capacity as well as improved backup functionality.

Customer feedback regarding the new technology options has been overwhelmingly positive.

Early VPP customers who had a Sunverge ESS installed have been offered a free upgrade to their choice of

new technology.

Customers who had paid a deposit for the previous technology, and had been waiting to have a system

installed, have now been contacted and progressively requoted on their choice of Tesla or LG Chem/Solar

Edge product.

2.2 Technology selection process

The technology selection process for new ESS technology involved four key stages:

Market survey

Based on learnings from the first phase of the program about ESS features that support consumer value and

both network support and wholesale participation use cases, an extensive market survey exercise was

undertaken focusing on a review of technology capability and specifications, including:

• compliance with Australian and international safety and performance standards.

• vertical integration of inverter control through an Application Program Interface (API)

• availability and distribution networks in Australian market

• vendor experience and support networks to deliver a ‘best in market’ customer experience

• ESS size, weight and complexity of installation

• quality of manufacture and warranties offered including performance

• physical integration of componentry such as the inverter backup capability

• three-phase site monitoring (including three-phase solar inverters)

• cost


Technical due diligence

AGL installed test units from shortlisted vendors in its energy storage lab at 699 Bourke Street in Melbourne

and in a limited number of residential homes. The purpose of this testing was to validate ESS performance

and trial the hardware installation. Key focuses included:

• vendor supplied performance data

• installation process

• safety response validation to faults conditions:

− loss of grid

− response to high and low grid voltage

− response to power limitations

• performance baselining:

− confirmation of charge/discharge power capabilities

− confirmation of response to a range of loads and changes in load on a range fo timescales

− ESS performance during low states of charge and low rates of discharge

− verification of changeover to backup power in the event of loss of grid.

Final selection

AGL’s final selection was the Tesla Powerwall 2 and the LG Chem RESU 10H - SolarEdge combination.

Multiple vendors were selected to mitigate the risk of supplier stock constraints, to demonstrate the

orchestration of multiple hardware brands and technologies, and to support customer choice.

2.3 Technology Announcement

AGL announced the updated VPP technology in targeted communications by email, supported by website

landing pages, with the objective of:

• announcing the new technology itself and communicating the opportunity to upgrade

• providing customers with sufficient information about the new product and clarity around the next steps to

upgrade their ESS

• assisting with customer education about ESS technology available.


VPP landing page – new technology announcement Tesla Powerwall 2

VPP landing page – new technology announcement LG Chem RESU 10H

Around 97% of customers clicked through to review the product-specific landing pages to learn more.

Customers spent a significant amount of time on each product page, investigating the new options and next

steps. Following the targeted email, AGL’s SA-VPP sales team made contact by phone to each customer,

discussed their product and installation options with them, and arranged for an ESS upgrade quotes to be

issued to the customer, if they chose to upgrade.


2.4 Sales

The addition of new battery technology to the SA-VPP program has been welcomed by customers and

‘requoting’ sales conversations have been less complex and considerably shorter than the initial sales calls.

Brand recognition amongst customers is very high for both products.

The dedicated VPP sales team have found that many customers have a good working knowledge of ESS

features and functionality and many already have a technology preference by the time the customer is called

to discuss their product preference. Of the more than 700 battery sales made to date, more than 500 have

been requoted and sold based on the new ESS technology.

2.5 Delivery and Installation Capacity

AGL has now trained seven installation teams in the safe delivery, installation and commissioning of both

technology offerings. These teams are both CEC and ESS manufacturer-accredited.

An additional three teams, bringing the total to ten, will be brought on line shortly.

2.6 Sales and installation metrics

As at 31 May 2018, AGL had completed 312 installations and is currently installing at a rate of 15 - 20 per

week, with a plan to increase this to 25 per week during June. Of the 312 ESS in the field at the end of May

89 were new technology installs, comprising a combination of greenfield sites and sites where the previous

technology ESS had been removed.


Overall project performance against milestones has been as shown below:

Stage Installation

target

Completion

date

Completed

installations

Achievement

Date

Milestone 1 65 28 Feb 2017 150 30 Apr 2017

Milestone 2 300 31 May 2018 312 31 May 2018

Milestone 3 1,000 (min. 950) 31 Dec 2018

2.7 VPP Performance

Over the summer of 2017/2018 the program had approximately 280 Sunverge systems in the field. The

fleet’s normal operating methodology is to reduce customer bills by maximising solar self-consumption.

This it does by targeting a zero-energy draw from the grid, charging the ESS when solar production

exceeds household load, and discharging the ESS when household load exceeds solar production. The

aim is to have households with, on average, a lower peak demand and lower export peaks.

2.7.1 VPP Orchestration Events

AGL undertook a ten-day orchestration trial between 19 - 29 of December 2017. During the trial, ESS were

directed to discharge with a site export target of 3 kW between 5pm – 6pm to demonstrate the fleet’s ability

to reduce the peak evening load on the grid from participating households.

The fleet’s ability to meet the targeted export value varied depending on factors including the weather and

customer load profile. These factors impact both the starting state of charge of the ESS fleet and the typical

household load during dispatch.

Beginning on 19 December 2017, AGL performed ten days in a row of one-hour orchestration events between

5pm – 6pm, the time of peak evening load, where site export targets were set to 3kW. On a sunny day, the

ESS were fully charged and the sites would typically have been exporting 1 – 2kW of solar before any

additional ESS export. This meant that the ESS needed to contribute an extra 1 – 2kW of power over the

hour.


2.7.2 Fleet Dispatch Details and Results

Figure 1 – energy export on dispatch days

The figure above is a plot of the energy exported from the sites for the ten events. Total Export is the total

site energy that was exported, Battery Export is the ESS energy exported from the site, and Solar Export is

the solar energy exported from the site. The purple line is the average state of charge (SOC) of the fleet at

the beginning of the dispatch. Note how closely it follows the total site export.

The graphs below display a three-hour time window (4pm – 7pm) centred around one-hour dispatches. The

top chart is how the mean behind-the-meter load is being served by the energy sources, green is load served

by the PV, orange is load served by the ESS, blue is load served by the grid.

The middle chart shows which energy source is exporting from the site. The PV, in green, slowly decreases

over the hour. The orange is the amount of ESS export required to make up the 3kW site export target. The

bottom chart is the average SOC and grid voltage of the fleet over the three hours.

0

10

20

30

40

50

60

70

80

90

100

0

100

200

300

400

500

600

700

800

900

19 20 21 22 23 24 25 26 27 28

SO

C (

%)

Energ

y E

xport

(kW

h)

Date of December 2017

December 2017 Export Events

Battery Export (kWh) Solar Export (kWh) Total Export (kWh) SOC Begin (%)


Figure 2 – dispatch performance on a sunny day

The Figure above shows a sunny day with solar generation before the evening export. The ESS were at or

near full SOC and there was around 2kW of solar being exported just before 5:00pm. At the start of the hour

the ESS added around 1kW of export, and ended the hour adding close to 2kW of export.


Figure 3 – dispatch performance on a cloudy day

The graphs above show a day with poor solar generation throughout the day and the ESS began the 5pm

dispatch not fully charged, averaging 60% SOC. There was little solar export at the 5pm start time, therefore

to meet the 3kW target the ESS provides almost all the site export, with limited contribution from solar at that

time.

2.7.3 Key Learnings

There were several interesting lessons from these trials, the key one being that it wasn’t possible to maintain

a consistent 3kW export target over the course of the one-hour dispatches. The reasons for this include high

grid voltage and SOC at a level where ESS stop exporting – but in addition sites with high loads did not meet

the export target.


Figure 4 – export power flow limited by load

The figure above illustrates why not all sites could reach a 3kW site export target. If there is behind the meter

load after being served by PV of greater than 2kW, the battery inverter capacity of 5kW limits the power

system’s ability to export the full 3kW.

Whilst the need for DNSP site export limits is understood, this trial raises a further question regarding battery

inverter capacity limits. For example, a 10kW battery inverter would be better at providing 5kW of site export

as it could better serve behind the meter loads while still having export capacity.

2.7.4 Day-to-Day Operation

In aggregate, the VPP significantly reduces the peaks in the loads drawn from the grid under the daily solar

self-consumption operation. In the graphs below, the blue curve denotes the gross loads before solar

(hypothetical), orange denotes the net loads if there was only solar (hypothetical), and then the green curves

are the net loads with the ESS included (actual results).


Figure 5 – average fleet shape across the year

The graph above shows the changing load shapes for March 2017, June 2017, September 2017, December

2017 and March 2018.

The blue curve is the gross load that the households use, which is what the grid would have had to deliver

in the absence of any installed solar or ESS.

The orange curve is the load that would have occurred if the site had a solar system only installed. Through

midday the load is significantly reduced, to almost zero (0.1kW average across the entire fleet, 0.2kW in June

with the least amount of solar generation). However, the peak load drawn by the sites only slightly reduces.

For example, it reduces from about 1kW to 0.8kW in March, and from slightly over 0.9kW to 0.75kW in

December.

In June solar does not reduce the peak load at all, which stays slightly above 1.3kW.

The green curves show that adding the ESS significantly flattens the load seen by the grid.

December and March have aggregate load values of around 0.2kW. In June, with the reduced solar

generation, the ESS are less charged and less able to shift load. The June peak of 0.8kW, with ESS included,

exceeds the summer peak.


Figure 6 – average fleet shape in spring and summer

The graph above shows the same chart for spring and summer. January has the significantly highest gross

load (blue) peaking at close to 1.3kW, though after the aggregated solar and ESS self-consumption, the peak

load comes down to around 0.4kW.

These data provide evidence, as expected, of the benefit of ESS to significantly reduce the peaks in the

loads drawn from the grid under the daily solar self-consumption operation, this is true compared to the

counterfactual of these sites with no solar, and solar only.

2.7.5 Reduced Solar Export

There has also been a significant reduction in the solar exports to the grid. In the figures below the blue

curve denotes the gross solar generation, orange curves denote the solar export that would have occurred

with the site load but without the ESS, and the green curves show what happened in aggregate with the ESS

performing solar self-consumption.


Figure 7 – reduced solar export due to ESS

The reduction in solar exported to the grid due to the ESS is clear from the figure above. The addition of the

ESS reduced the peak export by around 0.5kW, bringing the peak exports in December 2017 and March

2018 down from around 2.4kW to 1.9kW. For sites that discharge their ESS each night, the larger ESS being

installed later in our VPP program will reduce these exports even further.

2.7.6 VPP Performance Summary

The data shows that the ESS fleet has a high level of solar self-consumption, has reduced peak solar exports

during the day and has flattened the daily load profiles, especially through summer.

2.8 Summary of Project Media

Most media for the VPP was concentrated around the initial announcement and commencement of the

project (see Stage 1 milestone report), however there have been some stories in the media relating to the

project in Stage 2. These include:

• The Advertiser – Major Projects Conference attendees to get AGL power plant update in Adelaide


http://www.adelaidenow.com.au/business/sa-business-journal/major-projects-conference-attendees-to-

get-agl-power-plant-update-in-adelaide/news-story/dacd0a814f459dd9c6a10abaf6f623f1

• ABC News – AGL suspends household battery installations for Adelaide's cutting-edge Virtual Power

Plant, 7 September 2017

http://www.abc.net.au/news/2017-09-07/battery-scheme-under-the-spotlight/8883330

• Renew Economy – AGL hits pause on virtual power plant in technology “rethink”, 29 August 2018

https://reneweconomy.com.au/agl-hits-pause-on-virtual-power-plant-in-technology-rethink-57487/

• Renew Economy – AGL switches to Tesla and LG Chem for virtual power plant, 13 March 2018

https://reneweconomy.com.au/agl-switches-to-tesla-and-lg-chem-for-virtual-power-plant-77717/

• The Conversation – The unholy alliance that explains why renewable energy is trouncing nuclear, 20

March 2018

https://theconversation.com/the-unholy-alliance-that-explains-why-renewable-energy-is-trouncing-

nuclear-93519

http://www.adelaidenow.com.au/business/sa-business-journal/major-projects-conference-attendees-to-get-agl-power-plant-update-in-adelaide/news-story/dacd0a814f459dd9c6a10abaf6f623f1

http://www.adelaidenow.com.au/business/sa-business-journal/major-projects-conference-attendees-to-get-agl-power-plant-update-in-adelaide/news-story/dacd0a814f459dd9c6a10abaf6f623f1

http://www.abc.net.au/news/2017-09-07/battery-scheme-under-the-spotlight/8883330

https://reneweconomy.com.au/agl-hits-pause-on-virtual-power-plant-in-technology-rethink-57487/

https://reneweconomy.com.au/agl-switches-to-tesla-and-lg-chem-for-virtual-power-plant-77717/

https://theconversation.com/the-unholy-alliance-that-explains-why-renewable-energy-is-trouncing-nuclear-93519

https://theconversation.com/the-unholy-alliance-that-explains-why-renewable-energy-is-trouncing-nuclear-93519

Documents

Virtual Power Plant in South Australia · VPP-SA Milestone 2 Public Report 15-06-18 V1.1.docx 3 1 Introduction 1.1 Project overview AGL’s Virtual Power Plant (VPP) remains a key